I am calculating/plotting a log trendline extract on the equation
y = 533371Ln(x) - 3E+06. Ln(x) is the period (extract) number, starting at
1 and ending at 288. There are 288 plotted points. The value for 533€¦ and
3E€¦ are derived from the trendline equation each month, when a new value (and
account balance) is added to the series. The plotted extracted values (y)
give me an exact duplicate of the Excel log trendline curve, and, a value
(forcast) for period 288, where the trendline crosses the right side of the
chart. The series looks like this, where cell L2 holds the 533 value, M2
holds the 3E value, and G* is the extract number. Works fine.

=(L2)*LN(G1)-M2
=(L2)*LN(G2)-M2
=(L2)*LN(G3)-M2
etc.
=(L2)*LN(G288)-M2

Now I want to do a second series, using a 2nd order poly trendline with the
equation y = 1.8485x2+641.06x (given by Excel). The extract dimension
remains 1 through 288, and the series is:

=1.8485*(G1)+641.06*(G1)
=1.8485*(G2)+641.06*(G2)
=1.8485*(G3)+641.06*(G3)
etc.
=1.8485*(G288)+641.06*(G288)

The final value (G288) is coming up 183,228, when it should be in the region
of 340,000, the approximate value where the Excel-plotted trendline crosses
the right side of the chart.

Where is the logic flaw?